Steve Eisman has sold his long-held Google position to cut his artificial intelligence (AI) exposure. The investor who shorted the 2008 housing market now holds cash, warning the whole market has become one AI bet.

He has not bought a replacement. Eisman says defensive stocks will not work, because investors either want AI or they want nothing.

Eisman Sold Google Near Its Record High

Speaking on CNBC’s Squawk Box, the former Neuberger Berman portfolio manager called the exit deliberate. He built his reputation shorting subprime mortgages at FrontPoint Partners.

“I sold my Google a couple of months ago. I’ve owned Google. I can’t even tell you how long I’ve owned Google, but I felt I wanted to reduce my exposure to AI,” Eisman said.

The timing looks good so far. Alphabet peaked at $408.61 on May 18, its record high. The stock closed at $319.74 on July 24. That is a drop of roughly 20% in about two months.

One session did much of the damage. Alphabet fell 7.1% on July 23, the day after Q2 earnings. The company had just raised 2026 capital spending guidance to a range of $195 billion to $205 billion.

Eisman did not rotate into safety. He explained why in one line.

“People either want to buy AI or they don’t want to buy AI, but they don’t want to shift out of it to buy Clorox,” he said.

The cash is still uncommitted. “I’m just sitting… I’ve got cash,” he said. He does not expect the AI debate to settle “within the next two weeks.”

Why Eisman Says the Market Is ‘One Trade’

His worry is concentration, not valuation.

“It’s all one trade. It’s literally one,” Eisman said.

He then showed his math on a standard portfolio.

“Even people who think they’re diversified because they own 60% stocks and 40% bonds are missing the fact that they’re actually not diversified… more than 50%… is tech and AI related. And of the 40% of bonds, most of the new issuance of bonds is AI related,” he said.

Do Eisman’s Numbers Hold Up?

The stock half broadly does. Information Technology was 37.19% of the S&P 500 on July 24, and Communication Services added 9.34%. That is 46.5% combined.

Add Amazon and Tesla, which sit in Consumer Discretionary, and the figure reaches 51.5%. So his “more than 50%” works, but only on a generous definition.

The concentration itself is not in doubt. The 10 largest constituents make up 36.85% of the index.

The bond half is weaker. High Technology made up 14.2% of US corporate bond issuance in the second quarter, according to SIFMA. Financials led with 46.4%. AI is not “most” of new issuance.

His underlying point still stands, and official data makes it better. The Bank of England reported this month that five AI hyperscalers held just 3% of outstanding US investment-grade debt at the end of 2025, yet accounted for over 15% of this year’s issuance by early May.

The high-yield shift is sharper. Those issuers took 41% of non-refinancing US high-yield issuance this year, from a 1% index weight.

The deal sizes explain the speed. Amazon priced $37 billion of notes on March 10, the largest of these deals, per its SEC filing. Meta raised $30 billion last October and another $25 billion in April.

One caveat sits in the paperwork. The filings state proceeds go to general corporate purposes, so none of this debt is formally earmarked for AI.

Is a Correction Coming if AI Fails?

Asked what happens if AI fails commercially, Eisman was blunt. “I think we have a big correction,” he said. He would not size it.

“What… scares me is that it’s all one trade. So it better succeed,” he added.

Central banks have flagged the same pipe. The Bank for International Settlements warned in June that fixed income is “one obvious vulnerability” if hyperscalers slow capital spending.

What It Means for Crypto

Crypto sits in the same risk bucket. Bitcoin (BTC) trades near $64,980 and is down about 45% over the past year.

The link showed up in June, when a Big Tech selloff dragged Bitcoin lower. Retail flows have favored semiconductor ETFs over crypto funds this year.

Others see the same overlap. Chinese hedge funds have started trimming AI winners in a visible rotation, and one 2008 bubble forecaster has warned of a 70% drawdown.

Where the Thesis Breaks Down

Eisman is not calling a crash. He said he would not short this market, and he expects the technology to work.

“It’s going to be… something really good. That doesn’t mean that everybody succeeds,” he said.

That gap defines the risk. AI can succeed as a technology while the trade built around it still unwinds.

The near term will test him fast. Microsoft and Meta report earnings on July 29, and Amazon follows on July 30. Three more capex updates land inside 72 hours.

Source: BeInCrypto